10 Pb/sec? Re: IB vs Myrinet

[Joseph Landman]

On Tue, 2003-11-04 at 19:27, Keith D. Underwood wrote:
> > Say you can run a wire/fiber/whathaveyou at 10 Gb/sec (working in powers of 
> > 10, here, and 100 Gb/sec is just too hard to envision..)
> > 
> > 10 Pb/sec would be a million times faster (Peta = 1000 Tera = 1000,000 
> > Giga, right?)
> 
> Um: http://pr.fujitsu.com/en/news/2000/09/25.html
> or: http://www.siemens.com/page/1,3771,228164-1-999_4_0-0-,00.html
> 
> Multiple Tb/s on a single fiber...

<obHumor>
Get yer lambda's here... red hot... lots of them... 10000 to a fiber ...
</obHumor>

Think of this like the cable TV coax coming into your house
(theoretically if need be).  One wire, lots of bandwidth.  I have heard
(e.g. bad memory) that TV signals require 6MHz of bandwidth, so
"hundreds" of TV stations require somewhat less than a GHz in
bandwidth.  Same effect, using different 1/lambda's for each "channel".


> Of course, I don't have any idea what they are going to do with the data
> when they get it there either ;-) (i.e. issues with memory, buses,
> processor speeds, etc.) Or, for that matter, how they are going to get
> that data out of any silicon chip...   But, I assume these are planned
> for after we are doing all optical computing, right?  ;-) 

The entropy generation rate must be huge with all that data going into
the bit bucket, 
http://www.nature.com/nature/journal/v406/n6799/box/4061047a0_bx1.html  
unless of course a fast enough receiver can do something about this...

k(B) ln 2 for each bit "erased", so something like ((10**15) k(B) ln
2)/second, if you consider that bits dropped on the floor are erased. 
On the order of 10**(-8) W/K.  I defer to practicing physicists on the
interpretation (as it is not my field).


-- 
Joseph Landman, Ph.D
Scalable Informatics LLC
email: landman at scalableinformatics.com
  web: http://scalableinformatics.com
phone: +1 734 612 4615


_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf